
The US Just Banned Anthropic's Best Models and Nobody Can Explain Why
Export controls hit Fable and Mythos this week in a move that's left cybersecurity professionals furious and the rest of us scratching our heads.
Crédit photo: Image via The Verge — AI. Used under fair use for news commentary. · source
Most of the coverage this week focused on Anthropic's scramble to get back online. Fair enough. But the part that's getting buried is the bit that should be making everyone uncomfortable: nobody, including apparently Anthropic itself, knows exactly what legal authority the Trump administration used to do this.
Let me back up. Here's what actually happened. The administration ordered Anthropic to cut off all foreign nationals from its two newest models, Fable 5 and Mythos 5. Not just overseas users. Foreign nationals inside the US too. And, for a stretch, Anthropic's own employees. The company had to block access to both models entirely while it sorted out how to comply. The Verge reported that Anthropic's public statement only said the government cited "national security authorities" as justification, which is basically the regulatory equivalent of "because we said so."
One observer quoted in that piece put it plainly: "To my knowledge, this is the first time US export controls have been used to control access to an AI model in this way." That's not a small thing to bury in paragraph eight.
Now, I'll be honest. I spent twelve years at Kuka working on industrial control systems, not AI policy. Export controls were always someone else's headache, usually the legal team's, and even then the rules around dual-use technology were complicated enough that we'd sometimes spend weeks figuring out whether a specific actuator configuration needed an export license before it shipped to a customer in Southeast Asia. The rules existed for real reasons. I'm not dismissing that.
But those rules had a legal basis you could point to. The Commerce Department's Export Administration Regulations, the munitions list, specific classifications. You could argue about whether something belonged on the list, but at least there was a list. What we apparently have here is an order with no publicly stated legal foundation, applied retroactively to a software product that was already deployed, affecting users who were already inside the country. That's a different kind of thing.
À lire aussi
More in AI Models
Chipmakers swung wildly this week, from a Tuesday 'chip-wreck' to a Micron-led surge after hours. What's actually going on with AI's hardware backbone?
Sarah Williams · 26 Jun · 5 min
The original Creator Studio was shut down in 2023. Now it's back, rebuilt around an AI assistant that promises to grow your audience and reply to comments in your voice.
Sarah Williams · 26 Jun · 5 min
At its annual Config conference, Figma announced coding layers, AI-generated motion graphics, and a reimagined canvas that blurs the line between design and full-stack development.
Sarah Williams · 26 Jun · 5 min
Everyone talks about chips and models. The memory bottleneck is the part of the AI buildout that keeps getting underestimated, and Micron's latest earnings make that case hard to ignore.



